Functional changes in transcriptomes of the prefrontal

Functional changes in transcriptomes of

the prefrontal cortex and hippocampus in a mouse

model of anxiety

Dezso P. Virok1*, Zoltán Kis2*, Viktor Szegedi1, Gábor Juhász3, ÁgnesZvara2, Jr., Géza Müller4, György Lévay4, László G. Hársing5, RóbertRajkó6, Botond Penke1,3, Zoltán Janka7, Tamás Janáky3, László G. Puskás2

�Institute for Plant Genomics, Human Biotechnology and Bioenergy (BAY-GEN), 6726 Szeged Derkovits fasor 2.,Hungary

�Laboratory for Functional Genomics, Biological Research Center, Hungarian Academy of Sciences, 6726 SzegedTemesvári krt. 62., Hungary

�Department of Medical Chemistry, University of Szeged, 6720 Szeged Dóm tér 8., Hungary

�EGIS Nyrt., 1106 Budapest Keresztúri út 30-38. Budapest, Hungary

�Institute of Pharmacology and Pharmacotherapy Semmelweis University, 1085 Budapest VIII. Üllõi út 26., Hungary

�Faculty of Engineering, University of Szeged, 6725 Szeged, Moszkvai krt. 5-7., Hungary

�Department of Psychiatry, University of Szeged, 6725 Szeged, Semmelweis u. 6., Hungary

Correspondence: Dezso P. Virok, e-mail: [email protected]

Abstract:

Anxiety is a multi-etiology disorder influenced by both genetic background and environment. To study the impact of a genetic pre-

disposition, we developed a novel mouse model of anxiety using a combination of crossbreeding and behavioral selection. Compari-

son of the transcriptomes from the prefrontal cortex and hippocampus of anxious and control mice revealed that the numbers of

significantly up- and down-regulated genes were modest, comprising approximately 2% of the tested genes. Functional analysis of

the significantly altered gene sets showed that functional groups such as nervous system development, behavior, glial cell differen-

tiation and synaptic transmission were significantly enriched among the up-regulated genes, whereas functional groups such as po-

tassium ion transport, Wnt signaling and neuropeptidergic signaling were significantly enriched among the down-regulated genes.

Many of the identified genes and functional groups have been previously linked to the molecular biology of anxiety, while several

others, such as transthyretin, vasoactive intestinal polypeptide and various potassium ion channels, are novel or not as well described

in this context. Supporting the gene expression data, we also found increased excitability in the hippocampi of anxious mice, which

can be a phenotypic result of decreased potassium channel density. Our transcriptome screen showed that the initiation and/or effect

of anxiety involve multiple pathways and cellular processes. The identified novel genes and pathways could be involved in the mo-

lecular pathogenesis of anxiety and provide potential targets for further drug development.

Key words:

anxiety, gene expression, microarray, mouse model, hippocampus, prefrontal cortex

348 ��

��

��

� ��

��

�� !�"��

��"�� #!�$� � �� !��!��

* These authors contributed equally

Abbreviations: ACE – angiotensin convertase enzyme, ACSF

– artificial cerebrospinal fluid, ADCYAP1 – adenylate cyclase-

activating polypeptide 1, AX – anxious, C – control, cAMP –

cyclic adenosine monophosphate, CARTPT – cocaine and

amphetamine-regulated transcript, CGRP – calcitonin-gene re-

lated peptide, CREB – cAMP response element binding, fEPSP

– field excitatory post synaptic potential, GO – gene ontology,

GRIA1 – glutamate receptor, ionotropic 1, GRM1 – glutamate

receptor, metabotropic 1, GRP – gastrin releasing peptide,

KCNB1 – potassium voltage-gated channel, Shab-related sub-

family, member 1, KCNJ4 – potassium inwardly-rectifying

channel, subfamily J, member 4, KCNQ5 – potassium channel,

voltage gated, KQT-like subfamily, member 5, MAPK – mito-

gen-activated protein kinase, MEA – multi electrode array,

NPY – neuropeptide Y, NT – neurotensin, PDYN – prodynor-

phin, PENK – proencephalin, PCx – prefrontal cortex, QPCR –

quantitative real-time PCR, TBS – theta-burst stimulation, VIP

– vasoactive intestinal polypeptide

Introduction

Anxiety spectrum disorders exhibit a high prevalence

and have a serious medical and socioeconomic impact

in developed countries [2, 21]. The disease is multi-

factorial, requiring an interplay between genetic and

environmental factors. The genetic background of

anxiety has been demonstrated to be polygenic, most

likely involving genes from different cellular pro-

cesses and pathways [53, 58]. The identification of

the individual genes related to anxiety is a challenging

task due to the complex gene interactions that finally

lead to the disease. This complexity is mirrored by the

fact that anxiolytic drugs are effective only in a cer-

tain proportion of patients; approximately 40% of

those treated have limited or no response to these

therapies [16]. Above the molecular organizational

level, the anxiety-related neural circuits are also com-

plex, involving the prefrontal cortex (PCx), thalamus,

the hypothalamic-pituitary-adrenal axis, amygdala

and hippocampus. While the amygdala is central in

the formation of emotional responses, especially fear,

two other regions, the hippocampus and ventromedial

and dorsomedial PCx, are preferentially involved in

the contextualization of the proposed threat and the

learning of aversive behavior [3].

The genetic complexity of the disease, the inherent

biological variability of human patients, and the diffi-

culties in collecting human brain samples make it dif-

ficult to study the molecular mechanisms underlying

anxiety. Although animals can only mimic certain do-

mains of the anxiety-related complex behavioral phe-

notypes seen in humans, the availability of animal tis-

sues and the low genetic variability of inbred mouse

or rat strains make animal models attractive. The se-

lection and crossbreeding of extreme anxiety-related

phenotypes could result in a relatively homogenous

population in which the anxiety-related behavior is

a consequence of the same or similar alterations in the

molecular and neuronal networks of the brain. Ac-

cordingly, we applied a behavioral selection strategy

to construct an animal model in which the mice with

extreme anxiety phenotypes were selected and cross-

bred. Selection was made on the basis of anticipatory

anxiety behavior during the handling procedure.

When mice were moved to another cage, some indi-

viduals always ‘volunteered’ earlier, while others al-

ways moved away from the experimenter. Animals

were separated into two groups, early movers (non-

anxious, control) and late movers (anxious, AX), and

were inbred. After 35 generations, the selected AX

breed displayed significant anxiety relative to the

non-anxious controls in different well-established

anxiety tests, such as the open field, elevated plus

maze and light dark tests [54]. The complex nature of

anxiety requires complex, generally open systems in

order to explore the underlying molecular networks.

DNA microarray technology has been proven to be

highly efficient in describing molecular networks in

complex biological processes, such as development,

immunity, cancer progression [35] and various psy-

chiatric disorders [37]. We used DNA microarrays to

study the anxiety related transcriptome changes in

two relevant brain regions, the PCx and hippocampus,

of AX and control mice. We found several genes and

gene networks that were differentially expressed in

one or both of the studied regions. Several of the iden-

tified genes have been shown previously to be in-

volved in anxiety, while others are new and could pro-

vide valuable targets for future drug discovery.

Materials and Methods

Animal care and handling

Individuals of a non-commercially available inbred

mouse strain (MG15) were maintained at EGIS Phar-

maceuticals Co. (Budapest, Hungary). Mice from the

MG15 strain were crossed with NMRI mice (Charles

River Ltd., Budapest, Hungary), resulting a new

�� 349

Brain transcriptomes in a mouse model of anxiety��

strain. The behavioral selection was performed on in-

dividuals of this strain as described previously [54].

Experiments were performed on 12 AX and 12 con-

trol male mice 15–17-weeks old. The animals were

kept and the experiments were conducted in confor-

mity with Council Directive 86/609/EEC, the Hun-

garian Act of Animal Care and Experimentation

(1998, XXVIII) and local regulations for the care and

use of animals for research.

RNA preparation, amplification and labeling

Total RNA was purified from each mouse sample (PCx

and hippocampus tissue from each mouse) using an

RNA purification kit (Macherey Nagel, Düren, Ger-

many) according to the manufacturer’s instructions. At

a final concentration of 0.8 U/µl, an RNase inhibitor

(Fermentas, Lithuania) was added to the samples. RNA

quantity was determined using NanoDrop 3.1.0. RNA

samples were stored at –80°C before used. A sample of

the total RNA (1 µg) was amplified with the AminoAl-

lyl MessageAmpTM II aRNA Amplification Kit (Am-

bion, USA) according to the manufacturer’s instruc-

tions. An aliquot of the aminoallyl-modified amplified

RNA (aRNA; 6 µg) was labeled with either Cy5 or

Cy3 dye according to the manufacturer’s instructions

(Ambion, USA) then purified using RNA purification

columns from Macherey Nagel (Düren, Germany). The

dye incorporation rate and the labeled aRNA concen-

tration were detected using NanoDrop 3.1.0. The incor-

poration rate of the samples was 30–60 dye molecules

per 1000 nucleotides.

Microarray hybridization and data analysis

The mouse oligonucleotide microarray (8 × 15,000

format) from Agilent Technologies Inc. (Santa Clara,

CA, USA) was used to determine gene expression

changes. The hybridization was performed with the

Agilent Gene Expression Hybridization kit according

to the manufacturer’s instructions. Each array was

scanned at 543 nm (for Cy3 labeling) or at 633 nm

(for Cy5 labeling) in the Agilent Scanner using the

Extended Dynamic Range function of the ScanArray

software (10% and 100% photomultiplier values).

Twelve different samples from either the hippocam-

pus or the PCx (AX and control) were pooled into

four groups (three RNA samples each) and regarded

as biological replicas. In order to eliminate any distor-

tion caused by the different fluorescent dyes, parallel

dye-swap experiments were carried out from two

pools and regarded as technical replicas. Raw data

were analyzed with the Feature Extraction software

from Agilent Technologies (ver. 9.5.1.1.) and the

two-color gene expression protocol (GE2_v5_95_

Feb07). All the ratios were normalized using the

Lowess normalization method. The two-tailed two-

sample unequal variance Student’s t-test with a two-

fold change cutoff was applied to determine which

genes showed significantly altered expression levels

(GeneSpring GX 10.0.2.). The MIAME formatted mi-

croarray data can be found in the ArrayExpress Ar-

chive (http://www.ebi.ac.uk/microarray-as/ae/) with

the Accession Number: E-MEXP-2603. For func-

tional analysis of the altered genes, Gene Ontology-

based functional grouping was applied using the

DAVID web-based knowledge database [51]. The

Pathway Studio 6.0 (Rockville, MD, USA) and the

STRING web-based software [27] were used for path-

way and network analysis.

Quantitative real-time PCR (QPCR)

In order to validate the microarray data, QPCR was

performed on a RotorGene 3000 (Corbett Research,

Sydney, Australia) using gene-specific primers and

the SybrGreen protocol (Roche Applied Science,

Mannheim, Germany) as described previously [42].

Melting temperature analysis was done after each re-

action to check the quality of the reaction. Relative

expression ratios were calculated as normalized ratios

to the mouse �-actin gene. A control sample without

template was used for each PCR run to check for di-

mer formation in the primers. The final relative gene

expression ratios were calculated as �-� Ct values.

Electrophysiology

Using standard procedures, 350-µm-thick transverse

hippocampal slices were prepared from the brains of

2.5-month-old mice using a McIlwain tissue chopper

(Campden Instruments, Loughborough, UK). Slices

were incubated in standard artificial cerebrospinal

fluid (ACSF) at ambient temperature for 60 min with

95% O� – 5% CO�. The composition of the ACSF was

as follows (mM): NaCl, 130; KCl, 3.5; CaCl�, 2;

MgCl�, 2; NaH�PO�, 0.96; NaHCO�, 24; D-glucose, 10

(pH 7.4). Individual slices were transferred to a 3D-

multi-electrode array (MEA) chip with 60 tip-shaped,

60-µm-high electrodes spaced at 100 µm (Ayanda

350 ��

Biosystems, S.A., Lausanne, Switzerland). The slice

was continuously perfused with oxygenated ACSF

(1.5 ml/min at 34°C) throughout the recording ses-

sion. Data were recorded using a standard, commer-

cially available MEA setup (Multi Channel Systems

MCS GmbH, Reutlingen, Germany). The Schaffer-

collateral was stimulated by injecting a biphasic cur-

rent waveform (–100/+100 µs) through one selected

electrode at 0.033 Hz. Care was taken to place the

stimulating electrode in the same region from one

slice to the other. The peak-to-peak amplitudes of

field excitatory postsynaptic potentials (fEPSP) at the

stratum pyramidale and stratum radiatum of CA1

were analyzed. After a 30 min incubation period, the

threshold and the maximum of the stimulation inten-

sity required to evoke responses was determined. For

evoking responses, 30% of the maximal stimulation

intensity was used. Following a stable 15-min control

sequence, the stimulus intensity was continuously in-

creased from 0 to 70 µA in 10 µA steps (Input/Output

curve). Long-term potentiation (LTP) was induced us-

ing a theta-burst stimulation (TBS, 4 × 10 bursts of

4 pulses at 100 Hz) protocol applied at the maximum

stimulation intensity.

Results

Microarray analysis of the PCx and hippocam-

pus transcriptomes in AX and control mice

Twelve animals were sacrificed from the AX and con-

trol groups for transcriptome analysis. Total RNAs

were extracted from the hippocampus and PCx and

pooled so that each pool contained RNAs from 3 ani-

mals. For each brain region, four pooled samples from

AX mice were compared to four pooled samples from

control mice. A dye swap was also incorporated in the

experimental design. Two of the four hybridizations

were performed with Cy5 labeled AX samples, and

the other two were performed with Cy3 labeled AX

samples (Fig. 1). Approximately 15,000 genes from

the transcriptomes of the two brain regions were in-

vestigated using a custom made Agilent oligo mi-

croarray with 15,000 features. In order to obtain reli-

able microarray data, a two-step statistical analysis

was performed. A two-tailed two-sample unequal

variance Student’s t-test was used to determine the

p value that was used to find the significant gene

expression changes. Gene expression ratios with p value

< 0.05 and an AX/control ratio < –2 or > 2 were re-

garded as significant repression or over expression,

respectively, in gene activity. To assess the false dis-

covery rate, the microarray data were also investi-

gated with the Significance Analysis of Microarray

method (SAM) using the MultiExperiment Viewer

software [45]. The analysis showed that 94% and

99% of the hippocampus genes and 72% and 84% of

the PCx genes that were identified as significant by

our statistical analysis were also significantly changed

using SAM with a false discovery rate of 5% and

10%, respectively.

The numbers of significantly up regulated genes in

the AX mice compared to controls were similar in the

two brain regions. Sixty and 51 unique genes were

found to be up-regulated in the PCx and hippocam-

pus, respectively, and 97 genes were up-regulated in

�� 351


Fig. 1. Experimental design for mi-croarray analysis. Twelve animalseach were sacrificed from the AX andcontrol groups. Individual total RNAsamples from the PCx and the hippo-campus were pooled so that one poolcontained RNAs from three animals.For each brain region, four pooledsamples were tested from AX miceagainst four pooled samples from con-trol mice. To control dye bias, a dyeswap was also incorporated in theexperimental design. Two of the fourhybridizations were performed withCy5 (red)-labeled AX samples, andthe other two were performed with Cy3(green)-labeled AX samples (Fig. 1).HIPPO: hippocampus

both regions (Fig 2A). The numbers of down-

regulated genes were also similar in the two brain re-

gions: 41 and 47 unique genes were found to be

down-regulated in the PCx and hippocampus, respec-

tively, and 64 genes were repressed in both regions

(Fig. 2B). The pattern of gene expression was similar

in the two brain regions. We observed a relatively

high correlation ratio of 0.63 between the up-

regulated genes in the two brain regions, suggesting

that the direction and extent of up-regulation was not

region-specific in the AX brain (Fig. 2C). The pat-

terns of down-regulation in the brain regions were

less correlated, although the observed correlation co-

efficient of 0.31 showed that at least the direction of

down-regulation was similar in the two regions. None

of the genes were up-regulated in one region and

down-regulated in the other (Fig. 2D).

To validate the microarray data, the same RNA

samples that were used for microarray analysis were

used for an extensive QPCR screen. Altogether, 66

and 68 significantly changed genes detected by mi-

croarray were tested by QPCR in the PCx and hippo-

campus. Comparing the microarray and the QPCR

data, the results demonstrated a convincing 0.81 cor-

relation between the two methods. Moreover, of the

134 genes tested, only two (1.49% false positive)

demonstrated expression changes in opposite direc-

tions in the microarray and QPCR (Fig. 3 A, B).

Functional analysis of microarray data

In order to gain information about the general func-

tions of the altered genes, we performed a Gene On-

tology (GO) analysis, using the DAVID web-based

knowledgebase [51]. Generally, DAVID analyzes the

GO terms relating to the significantly altered genes,

identifies the terms that contain multiple genes, and

calculates a significance value for the observed en-

richment in comparison with a background gene set.

Because the directions of gene alterations were gener-

ally similar in the two regions and many altered genes

were shared, we analyzed the up- and down-regulated

genes in toto from both brain regions. The analysis re-

vealed that several of the GO biological function

groups that were significantly enriched in the AX

down-regulated gene set had been previously linked

to anxiety and/or mood disorders (Tab. 1). Among

others, these functional classes included potassium

ion transport, the neuropeptidergic signaling pathway,

nervous system development and the intracellular sig-

naling cascade. The down-regulated gene set was also

automatically mapped into previously described path-

ways using the Pathway Studio software. In accor-

dance with the GO analysis results, the pathway

analysis revealed that the down-regulated genes for

several neuropeptides and the corresponding adeny-

late cyclase–cAMP-MAPK, phosphatidyl-inositol–

352 ��

Fig. 2. Comparison of transcriptomechanges in the PCx and hippocam-pus. Venn diagram of the (A) signifi-cantly up-regulated and (B) down-regulated genes in the PCx and hippo-campus (HIPPO). Correlation betweenthe average log

�ratios of (C) induced

and (D) repressed genes in AX PCxand hippocampus vs. the control brainregions

�� 353


Fig. 3. QPCR validation of microarray data. The pools of RNA from the hippocampus and PCx of AX and control mice were used as templates in theQPCR reaction. The means ± SD of the AX/control ratio detected by microarray and QPCR are shown for the (A) hippocampus and (B) PCx samples

calcium–calmodulin, protein kinase A, and protein ki-

nase C downstream signaling members could be

grouped into a common pathway (Fig. 4). The up-

regulated AX gene set also contained GO biological

function classes possibly related to the observed AX

phenotype, including nervous system development,

behavior and synaptic transmission (Tab. 1).

The DAVID tissue expression database search showed

that the expression of 63 genes from the down-regulated

gene set had been previously linked to the brain, while 16

had been linked to the brain cortex. Similarly to the

down-regulated genes, the DAVID tissue expression data-

base search indicated that the expression of 40 genes

from the up-regulated gene set had been previously

linked to the brain, and 9 had been linked to the hippo-

campus. In addition to the QPCR data, the parallel ap-

pearance of central nervous system-related functional

categories and tissue expression patterns further vali-

dated our microarray experiments.

The STRING protein-protein interaction database

was used to explore all of the possible connections be-

tween the significantly altered genes (Fig. 5). The inter-

action analysis showed that 93 genes, 61% of the AX

down-regulated gene set, were connected to at least one

other gene. The sub-networks at least partially mirrored

the GO functional classes described above, such as the

Wnt signaling network, the neuropeptidergic network,

phosphatidyl-inositol signaling, the calcium signaling

networks and the potassium channel network. The

STRING analysis also revealed other sub-networks con-

taining immunity-related genes/proteins and Rho-Rac

genes involved in cytoskeleton reorganization and traf-

ficking. The STRING interaction analysis of the AX

up-regulated gene set showed that 75 genes, 36% of the

AX up-regulated genes, are connected to each other. The

extended STRING analysis identified sub-networks re-

lated to iodine-thyroxin metabolism, neuropeptidergic

signaling, the homeobox network, lipid metabolism and

inflammation.

Electrophysiology of hippocampal fEPSP and LTP

The down-regulation of the voltage-gated potassium

channel expression in the AX mice could influence

the properties of network excitation and may lead to

a net increase in neuronal excitability. In order to in-

vestigate this hypothesis, we used in vitro fEPSP re-

cordings from hippocampus slices generated with

a multi-electrode array setup (Fig. 6A). The excitabil-

ity was directly addressed by recording an I/O curve

from control and AX mice. The AX mice exhibited

significantly increased evoked fEPSPs from 30 µA

(500 ± 126 µV for AX mice, n = 4, vs. 325 ± 20 µV

for controls, n = 4; p � 0.05). This trend was also ob-

served for the stronger stimulation-evoked fEPSPs

(808 ± 172 µV vs. 472 ± 32 µV at 40 µA, 1057 ± 215

µV vs. 602 ± 44 µV at 50 µA, 1253 ± 252 µV vs. 731

± 57 µV at 60 µA and 1644 ± 532 µV vs. 956 ± 67 µV

at 70 µA, for the AX and control mice, respectively).

The observed enhanced excitation in AX mice

could lead to impaired LTP in the hippocampus.

Therefore, we investigated LTP. The level of theta-

burst stimulation (TBS)-evoked LTP was slightly

larger than the control 60 min after TBS (140 ± 3%

for the AX mice, n = 4 vs. 136 ± 4% for the control,

n = 4; p = 0.08). However, the difference was not sig-

nificant (Fig. 6B).

354 ��

Tab. 1. Gene Ontology analysis of microarray data. The significantlyup- and down-regulated genes in the hippocampus and/or PCx of AXmice were analyzed for significantly enriched GO functional groupsusing the DAVID web-based knowledge database. The significantlyenriched (p < 0.05) functional groups that could be related to anxietyand the corresponding p values are shown

Down-regulated genes in AX mice

Gene Ontology biological function term p value

metal ion transport 2.18 � 10–5

Wnt receptor signaling pathway 5.34 � 10–5

potassium ion transport 1.01 � 10–4

neuropeptidergic signaling pathway 5.74 � 10–4

intracellular signaling cascade 5.57 � 10–3

nervous system development 1.47 � 10–2

regulation of adenylate cyclase activity 1.76 � 10–2

synaptic transmission 2.73 � 10–2

Up-regulated genes in AX mice

Gene Ontology biological function term p value

nervous system development 3.58 � 10–5

behavior 2.06 � 10–4

neuron differentiation 1.90 � 10–3

regulation of neuron apoptosis 4.21 � 10–3

synaptic transmission 2.04 � 10–2

cell-cell signaling 2.07 � 10–2

neurite development 2.11 � 10–2

inorganic anion transport 3.18 � 10–2

glial cell differentiation 4.21 � 10–2

�� 355


Fig. 4. Molecular network analysis of microarray data. Down-regulation of the neuropeptidergic signaling network in AX mice (See the includedTable for gene symbols and level of down-regulation). The significantly repressed genes are highlighted in blue. The fold changes are aver-ages of four independent measurements in each tissue

356 ��

Fig. 5. Network analysis of microarray data. Interaction analysis of the up- and down-regulated AX gene sets. The significantly altered geneswere analyzed by using the STRING interaction database. The functional groups potentially related to anxiety are highlighted, and the genesrelating to the groups are shown with the fold change of up- and down-regulation in the AX hippocampus and PTX vs. controls. The foldchanges are averages of four independent measurements in each tissue

Discussion

Transcriptome changes in the AX prefrontal

cortex and hippocampus

To investigate the genetic background of anxiety, we

applied selection and further breeding of non-anxious

control and AX mice [54]. We hypothesized that the

resulting clearly distinct phenotypes could provide

a valuable basis to study the molecular networks cor-

related to anxiety [1].

Our transcriptome screen revealed that approximately

2% of the tested genes were significantly altered in at

least one of the brain regions of AX mice compared to

controls. The thorough QPCR validation of the mi-

croarray data showed a high correlation between the

two methods with a low rate of false positives. Func-

tional grouping of the significantly altered genes by

GO or STRING-based analysis revealed several func-

tional gene groups that could be associated with the

development of anxiety in both brain regions. Func-

tional groups, such as neuropeptides and develop-

ment-related groups, contained genes from both the

AX up- and down-regulated gene sets, while the po-

tassium ion channel group and the intracellular signal

transduction group contained predominantly down-

regulated AX genes.

Altered neuropeptides and signaling networks

One of the most prominent gene expression alteration

in both brain regions of the AX mice included numer-

ous neuropeptides and members of neuropeptidergic

signaling, primarily the phosphatidylinositol-calcium,

adenylate cyclase-cAMP-MAPK and calmodulin sig-

naling networks. It is interesting that while neuropep-

tide genes were both up-regulated and down-regu-

lated, the downstream signaling pathway members

were mainly down-regulated. The induced neuropep-

tides included angiotensinogen, calcitonin-gene re-

lated peptide, neurotensin, neurotrophin-3 and gastrin

releasing peptide. These up-regulated AX neuropep-

tides were previously shown to have cerebral loca-

tion- and concentration-dependent positive or nega-

tive effects on fear and anxiety responses [9, 12, 13,

19, 38, 50, 52, 55]. Among the down-regulated AX

neuropeptides, prodynorphin and proencephalin dis-

played an anxiolytic effect in different animal models

and anxiety tests [7, 29, 33, 59]. The connection be-

tween vasoactive intestinal polypeptide (VIP), down-

regulated in AX animals, and anxiety has not been

�� 357


Fig. 6. Input/Output curve and LTP recordings of AX and control hippocampi. (A) Input/Output curve recorded from hippocampi of AX and con-trol mice. Inserts show representative fEPSPs evoked by increasing stimulation intensity (red for AX and black for control mice). Calibrationbars are 25 ms and 1 mV. * Significant difference from the control at p � 0.05. (B) LTP in hippocampus of the AX and control mice. Inserts showrepresentative fEPSPs recorded at the indicated time points. Note the larger fEPSPs in the AX mouse. Calibration bars are 25 ms and 1 mV

well described, but the impact of VIP on behavior was

described previously. VIP has been associated with

learning- and memory-related behavior [20], social

behavior [26], and water and alcohol consumption

[32], all behavioral phenotypes that could be related

to anxiety. PACAP, a neuropeptide structurally related

to VIP, was down-regulated in AX and has also been

shown to have an impact on behavior [43, 56] and

memory retrieval [22]. Neuropeptide Y (NPY), an-

other neuropeptide down-regulated in AX, has been

clearly linked to anxiety behavior [8, 25, 44, 47] and

the central response to benzodiazepine treatment [24,

28]. It is interesting that there could be an inverse cor-

relation between the NPY concentration and the con-

centration of transthyretin, the most highly up-

regulated gene we found in AX mice. It has been pre-

viously shown that the NPY concentration was ele-

vated in the forebrain of transthyretin knock-out mice

[41], indicating a transthyretin-mediated NPY down-

regulation. The down-regulation of several signal

transduction pathways linked to neuropeptidergic net-

works were also observed in AX mice. Interestingly,

these networks seemed to be downstream effectors of

the transduction of VIP-PACAP and NPY. The VIP-

PACAP system uses the phosphatidylinositol–cal-

cium–calmodulin signaling pathways and the adeny-

late cyclase–cAMP-PKA-MAPK networks, while

NPY signals through the phosphatidylinositol–cal-

cium–calmodulin network. The phosphatidylinosi-

tol–calcium signaling converges into the calmodulin-

calmodulin kinase network, which targets, among

others, the cAMP response element binding (CREB)

transcription factor [15]. CREB is a common target of

the two signaling pathways. The adenylate cy-

clase–cAMP signaling also targets CREB through the

MAP-MAPK network [15, 60]. It is noteworthy that,

in addition to the involvement in the CREB pathway,

adenylate cyclase signaling has been associated with

the anxiety phenotype [17, 31, 49]. CREB is a central

transcription factor involved in different behavioral

processes and phenotypes, such as memory [34], in-

duction of LTP [10], synaptic plasticity [36] and de-

pression [4, 40, 57]. The connection between CREB

activity and anxiety has also been previously de-

scribed in different model systems [5, 6]. Overall, our

results indicate that the decreased CREB activation

might be involved in the initiation of anxiety in our

animal model.

Anxiety-related changes of ion channel and

receptor expression

One striking feature of the mRNA expression pattern

was the down-regulation of a large number of ion

channels, particularly potassium ion channels, in AX

mice. The significance of the enrichment of the metal

ion transport GO biological function category was p =

0.00002 with 17 repressed genes, while the potassium

channels category had a significance level of p = 0.0001

with 10 genes down-regulated. Among these chan-

nels, only a few have been characterized functionally

in relation to neural excitability and/or anxiety. The

AX down-regulated potassium channel, the voltage-

gated KQT-like subfamily member 5 (KCNQ5), is

a member of the KCNQ family that encodes the Kv7

potassium channel subunits. Kv7 channels are in-

volved in decreasing excitability [23], and Kv7 chan-

nel openers have been successfully applied in uncon-

ditioned and conditioned rodent models of anxiety

[30]. The down-regulated potassium voltage-gated

channel, Shab-related subfamily member 1 (KCNB1)

potassium channel, is a key mediator of the delayed

rectifier Kv currents regulating neuronal excitability

[39]. The AX down-regulated KCNJ4 potassium

channel has also been shown to play a role in decreas-

ing neural excitability in rat primary striatal cultures

[18]. In addition to the various potassium channels,

the delta subunit of the inhibitory GABA-A receptor

was also down regulated in the AX mice, which also

pointed to an imbalance between inhibition and exci-

tation. In support of the increased excitability hy-

pothesis, both the AMPA type ionotropic glutamate

receptor GRIA1 and the group I metabotropic gluta-

mate receptor GRM1 were found to be up-regulated

in both brain regions in AX mice. The involvement of

ionotropic AMPA glutamate receptors and group I

metabotropic glutamate receptors have also been de-

scribed in different models of anxiety [11, 14].

Altogether, the above-mentioned results high-

lighted the possibility that the neural excitability in

the PCx and hippocampus increased in the AX mice.

We addressed the increased neuronal excitability hy-

pothesis directly using electrophysiology in hippo-

campal slices. The excitability (fEPSP) was signifi-

cantly greater in the AX mice than in the control

strain, illustrating that the changes in the mRNA lev-

els were reflected in the phenotype. However, LTP

was intact, suggesting that higher cognitive functions

were not impaired in the AX mice. Our results indi-

358 ��

cate that, in addition to the already described in-

creased amygdalar excitability [46, 48], elevated hip-

pocampal excitability also correlates with the anxiety

phenotype.

Conclusions

In accordance with the multigenic nature of anxiety,

our transcription analysis of the AX hippocampus and

PCx clearly revealed that several genes with different

functions and from various pathways may be involved

in the generation of the observed phenotype. One of

the major questions is whether the observed alteration

of gene expression is a cause or an effect of anxiety.

Most likely, both mechanisms are involved in our

model. Pathways, such as the potassium channel net-

work, the neuropeptidergic system and the corre-

sponding signaling pathways, can have a causative

role in anxiety development, while others, such as the

inflammation-immunity related genes, may be a result

of the stress-related immune function modulation.

Our analysis served as a first screening approach. The

characterization of particular genes in the develop-

ment of anxiety requires more focused studies using

targeted loss of function or gain of function ap-

proaches.

Acknowledgments:

Dezso P. Virok and Viktor Szegedi were supported by the Ede TellerProgram of the National Office for Research and Technology(NAP_BIO_06). Agnes Zvara was supported by the János BolyaiFellowship of the Hungarian Academy of Sciences (BO/00381/07).This work was supported by the National Office for Research andTechnology (NKTH RET-08/2004) and TAMOP-4.2.2.-08/1-2008-0002. Botond Penke, Tamás Janáky, Gábor Juhász were supportedby Hungarian Scientific Research Fund (OTKA NK 73672) andthe European Commission (FP-7 201 159).

Conflict of interest

Authors declare no competing financial interests in relation tothe presented work.

References:

1. Andrade EC, Krueger DD, Nairn AC: Recent advances

in neuroproteomics. Curr Opin Mol Ther, 2007, 9,

270–281.

2. Ansseau M, Dierick M, Buntinkx F, Cnockaert P,

De Smedt J, Van Den Haute M, Vander Mijnsbrugge D:

High prevalence of mental disorders in primary care.

J Affect Disord, 2004, 78, 49–55.

3. Bannerman DM, Rawlins JN, McHugh SB, Deacon RM,

Yee BK, Bast T, Zhang WN et al.: Regional dissociations

within the hippocampus-memory and anxiety. Neurosci

Biobehav Rev, 2004, 28, 273–283.

4. Barad M, Bourtchouladze R, Winder DG, Golan H,

Kandel E: Rolipram, a type IV-specific phosphodies-

terase inhibitor, facilitates the establishment of long-

lasting long-term potentiation and improves memory.

Proc Natl Acad Sci USA, 1998, 95, 15020–15025.

5. Barrot M, Olivier JD, Perrotti LI, DiLeone RJ, Berton O,

Eisch AJ, Impey S et al.: CREB activity in the nucleus

accumbens shell controls gating of behavioral responses

to emotional stimuli. Proc Natl Acad Sci USA, 2002, 99,

11435–11440.

6. Barrot M, Wallace DL, Bolanos CA, Graham DL,

Perrotti LI, Neve RL, Chambliss H et al.: Regulation of

anxiety and initiation of sexual behavior by CREB in the

nucleus accumbens. Proc Natl Acad Sci USA, 2005, 102,

8357–8362.

7. Bilkei-Gorzo A, Racz I, Michel K, Zimmer A, Kling-

muller D: Behavioral phenotype of pre-proenkephalin-

deficient mice on diverse congenic backgrounds.

Psychopharmacology (Berl), 2004, 176, 343–352.

8. Broqua P, Wettstein JG, Rocher MN, Gauthier-Martin B,

Junien JL: Behavioral effects of neuropeptide Y receptor

agonists in the elevated plus-maze and fear-potentiated

startle procedures. Behav Pharmacol, 1995, 6, 215–222.

9. Catts VS, Al-Menhali N, Burne TH, Colditz MJ,

Coulson EJ: The p75 neurotrophin receptor regulates

hippocampal neurogenesis and related behaviours. Eur

J Neurosci, 2008, 28, 883–892.

10. Cha-Molstad H, Keller DM, Yochum GS, Impey S,

Goodman RH: Cell-type-specific binding of the tran-

scription factor CREB to the cAMP-response element.

Proc Natl Acad Sci USA, 2004, 101, 13572–13577.

11. Chojnacka-Wojcik E, Klodzinska A, Pilc A: Glutamate

receptor ligands as anxiolytics. Curr Opin Investig

Drugs, 2001, 2, 1112–1119.

12. Chourbaji S, Brandwein C, Vogt MA, Dormann C,

Hellweg R, Gass P: Nature vs. nurture: can enrichment

rescue the behavioural phenotype of BDNF heterozy-

gous mice? Behav Brain Res, 2008, 192, 254–258.

13. Costall B, Domeney AM, Gerrard PA, Horovitz ZP,

Kelly ME, Naylor RJ, Tomkins DM: Effects of captopril

and SQ29,852 on anxiety-related behaviours in rodent

and marmoset. Pharmacol Biochem Behav, 1990, 36,

13–20.

14. Das P, Lilly SM, Zerda R, Gunning WT, 3rd, Alvarez FJ,

Tietz EI: Increased AMPA receptor GluR1 subunit incor-

poration in rat hippocampal CA1 synapses during benzo-

diazepine withdrawal. J Comp Neurol, 2008, 511,

832–846.

15. Dash PK, Karl KA, Colicos MA, Prywes R, Kandel ER:

cAMP response element-binding protein is activated by

Ca��/calmodulin- as well as cAMP-dependent protein

kinase. Proc Natl Acad Sci USA, 1991, 88, 5061–5065.

16. de Mooij-van Malsen AJ, Olivier B, Kas MJ: Behav-

ioural genetics in mood and anxiety: a next step in find-

�� 359


ing novel pharmacological targets. Eur J Pharmacol,

2008, 585, 436–440.

17. de Mooij-van Malsen AJ, van Lith HA, Oppelaar H,

Hendriks J, de Wit M, Kostrzewa E, Breen G et al.:

Interspecies trait genetics reveals association of Adcy8

with mouse avoidance behavior and a human mood

disorder. Biol Psychiatry, 2009, 66, 1123–1130.

18. Falk T, Xie JY, Zhang S, Kennedy J, Bennett J, Yool AJ,

Sherman SJ: Over-expression of the potassium channel

Kir2.3 using the dopamine-1 receptor promoter selec-

tively inhibits striatal neurons. Neuroscience, 2008, 155,

114–127.

19. Georgiev V, Stancheva S, Kambourova T, Getova D:

Effect of angiotensin II on the Vogel conflict paradigm

and on the content of dopamine and noradrenaline in rat

brain. Acta Physiol Pharmacol Bulg, 1990, 16, 32–37.

20. Glowa JR, Panlilio LV, Brenneman DE, Gozes I, Fridkin

M, Hill JM: Learning impairment following intracerebral

administration of the HIV envelope protein gp120 or

a VIP antagonist. Brain Res, 1992, 570, 49–53.

21. Greenberg PE, Sisitsky T, Kessler RC, Finkelstein SN,

Berndt ER, Davidson JR, Ballenger JC, Fyer AJ: The

economic burden of anxiety disorders in the 1990s.

J Clin Psychiatry, 1999, 60, 427–435.

22. Hagino N: Performance of PAC1-R heterozygous mice

in memory tasks-II. J Mol Neurosci, 2008, 36, 208–219.

23. Hansen HH, Waroux O, Seutin V, Jentsch TJ, Aznar S,

Mikkelsen JD: Kv7 channels: interaction with dopamin-

ergic and serotonergic neurotransmission in the CNS.

J Physiol, 2008, 586, 1823–1832.

24. Heberlein A, Bleich S, Kornhuber J, Hillemacher T:

Neuroendocrine pathways in benzodiazepine depend-

ence: new targets for research and therapy. Hum Psycho-

pharmacol, 2008, 23, 171–181.

25. Heilig M, McLeod S, Koob GK, Britton KT:

Anxiolytic-like effect of neuropeptide Y (NPY), but not

other peptides in an operant conflict test. Regul Pept,

1992, 41, 61–69.

26. Hill JM, Hauser JM, Sheppard LM, Abebe D, Spivak-

Pohis I, Kushnir M, Deitch I, Gozes I: Blockage of VIP

during mouse embryogenesis modifies adult behavior

and results in permanent changes in brain chemistry.

J Mol Neurosci, 2007, 31, 183–200.

27. Jensen LJ, Kuhn M, Stark M, Chaffron S, Creevey C,

Muller J, Doerks T et al.: STRING 8 – a global view on

proteins and their functional interactions in 630 organ-

isms. Nucleic Acids Res, 2008.

28. Kask A, Rago L, Harro J: Anxiogenic-like effect of the

neuropeptide Y Y1 receptor antagonist BIBP3226: an-

tagonism with diazepam. Eur J Pharmacol, 1996, 317,

R3–4.

29. Konig M, Zimmer AM, Steiner H, Holmes PV, Crawley

JN, Brownstein MJ, Zimmer A: Pain responses, anxiety

and aggression in mice deficient in pre-proenkephalin.

Nature, 1996, 383, 535–538.

30. Korsgaard MP, Hartz BP, Brown WD, Ahring PK,

Strobaek D, Mirza NR: Anxiolytic effects of Maxipost

(BMS-204352) and retigabine via activation of neuronal

Kv7 channels. J Pharmacol Exp Ther, 2005, 314, 282–292.

31. Krishnan V, Graham A, Mazei-Robison MS, Lagace DC,

Kim KS, Birnbaum S, Eisch AJ et al.: Calcium-sensitive

adenylyl cyclases in depression and anxiety: behavioral

and biochemical consequences of isoform targeting. Biol

Psychiatry, 2008, 64, 336–343.

32. Kulkosky PJ, Doyle JS, Cook VI, Glazner GW, Foderaro

MA: Vasoactive intestinal peptide: behavioral effects in

the rat and hamster. Pharmacol Biochem Behav, 1989,

34, 387–393.

33. Kuzmin A, Madjid N, Terenius L, Ogren SO, Bakalkin

G: Big dynorphin, a prodynorphin-derived peptide pro-

duces NMDA receptor-mediated effects on memory,

anxiolytic-like and locomotor behavior in mice. Neuro-

psychopharmacology, 2006, 31, 1928–1937.

34. Kwok RP, Lundblad JR, Chrivia JC, Richards JP,

Bachinger HP, Brennan RG, Roberts SG et al.: Nuclear

protein CBP is a coactivator for the transcription factor

CREB. Nature, 1994, 370, 223–226.

35. Lobenhofer EK, Bushel PR, Afshari CA, Hamadeh HK:

Progress in the application of DNA microarrays. Environ

Health Perspect, 2001, 109, 881–891.

36. Marie H, Morishita W, Yu X, Calakos N, Malenka RC:

Generation of silent synapses by acute in vivo expression

of CaMKIV and CREB. Neuron, 2005, 45, 741–752.

37. Mirnics K, Levitt P, Lewis DA: Critical appraisal of

DNA microarrays in psychiatric genomics. Biol Psychia-

try, 2006, 60, 163–176.

38. Mountney C, Sillberg V, Kent P, Anisman H, Merali Z:

The role of gastrin-releasing peptide on conditioned fear:

differential cortical and amygdaloid responses in the rat.


39. Murakoshi H, Trimmer JS: Identification of the Kv2.1

K+ channel as a major component of the delayed recti-

fier K+ current in rat hippocampal neurons. J Neurosci,

1999, 19, 1728–1735.

40. Nibuya M, Nestler EJ, Duman RS: Chronic antidepres-

sant administration increases the expression of cAMP re-

sponse element binding protein (CREB) in rat hippocam-

pus. J Neurosci, 1996, 16, 2365–2372.

41. Nunes AF, Saraiva MJ, Sousa MM: Transthyretin knock-

outs are a new mouse model for increased neuropeptide

Y. FASEB J, 2006, 20, 166–168.

42. Puskas LG, Bereczki E, Santha M, Vigh L, Csanadi G,

Spener F, Ferdinandy P et al.: Cholesterol and cholesterol

plus DHA diet-induced gene expression and fatty acid

changes in mouse eye and brain. Biochimie, 2004, 86,

817–824.

43. Reglodi D, Kiss P, Tamas A, Lengvari I: The effects of

PACAP and PACAP antagonist on the neurobehavioral

development of newborn rats. Behav Brain Res, 2003,

140, 131–139.

44. Roy A, Pandey SC: The decreased cellular expression of

neuropeptide Y protein in rat brain structures during

ethanol withdrawal after chronic ethanol exposure.

Alcohol Clin Exp Res, 2002, 26, 796–803.

45. Saeed AI, Sharov V, White J, Li J, Liang W, Bhagabati

N, Braisted J et al.: TM4: a free, open-source system for

microarray data management and analysis. Biotech-

niques, 2003, 34, 374–378.

46. Sajdyk TJ, Shekhar A: Excitatory amino acid receptor

antagonists block the cardiovascular and anxiety re-

sponses elicited by �-aminobutyric acid� receptor block-

360 ��

ade in the basolateral amygdala of rats. J Pharmacol Exp

Ther, 1997, 283, 969–977.

47. Sajdyk TJ, Vandergriff MG, Gehlert DR: Amygdalar

neuropeptide Y Y1 receptors mediate the anxiolytic-like

actions of neuropeptide Y in the social interaction test.

Eur J Pharmacol, 1999, 368, 143–147.

48. Sanders SK, Shekhar A: Regulation of anxiety by

GABA� receptors in the rat amygdala. Pharmacol

Biochem Behav, 1995, 52, 701–706.

49. Schaefer ML, Wong ST, Wozniak DF, Muglia LM,

Liauw JA, Zhuo M, Nardi A et al.: Altered stress-

induced anxiety in adenylyl cyclase type VIII-deficient

mice. J Neurosci, 2000, 20, 4809–4820.

50. Shepherd J, Bill DJ, Dourish CT, Grewal SS,

McLenachan A, Stanhope KJ: Effects of the selective

angiotensin II receptor antagonists losartan and

PD123177 in animal models of anxiety and memory.


51. Sherman BT, Huang da W, Tan Q, Guo Y, Bour S, Liu D,

Stephens R et al.: DAVID Knowledgebase: a gene-

centered database integrating heterogeneous gene anno-

tation resources to facilitate high-throughput gene func-

tional analysis. BMC Bioinformatics, 2007, 8, 426.

52. Shilling PD, Feifel D: The neurotensin-1 receptor ago-

nist PD149163 blocks fear-potentiated startle. Pharmacol

Biochem Behav, 2008, 90, 748–752.

53. Smoller JW, Faraone SV: Genetics of anxiety disorders:

complexities and opportunities. Am J Med Genet C

Semin Med Genet, 2008, 148C, 85–88.

54. Szego EM, Janaky T, Szabo Z, Csorba A, Kompagne H,

Muller G, Levay G et al.: A mouse model of anxiety

molecularly characterized by altered protein networks in

the brain proteome. Eur Neuropsychopharmacol, 2010,

96–111.

55. Tashev R, Belcheva S, Milenov K, Belcheva I: Behav-

ioral effects of somatostatin microinjected into caudate

putamen. Neuropeptides, 2001, 35, 271–275.

56. Telegdy G, Kokavszky K: The action of pituitary adeny-

late cyclase activating polypeptide (PACAP) on passive

avoidance learning. The role of transmitters. Brain Res,

2000, 874, 194–199.

57. Wallace TL, Stellitano KE, Neve RL, Duman RS: Effects

of cyclic adenosine monophosphate response element

binding protein overexpression in the basolateral amyg-

dala on behavioral models of depression and anxiety.

Biol Psychiatry, 2004, 56, 151–160.

58. Weso³owska A: Potential role of the 5-HT� receptor in

depression and anxiety: an overview of preclinical data.

Pharmacol Rep, 2010, 62, 564–577.

59. Wittmann W, Schunk E, Rosskothen I, Gaburro S, Singe-

wald N, Herzog H, Schwarzer C: Prodynorphin-derived

peptides are critical modulators of anxiety and regulate

neurochemistry and corticosterone. Neuropsychophar-

macology, 2009, 34, 775–785.

60. Xing J, Ginty DD, Greenberg ME: Coupling of the RAS-

MAPK pathway to gene activation by RSK2, a growth

factor-regulated CREB kinase. Science, 1996, 273,

959–963.

Received: May 4, 2010; in the revised form: October 6, 2010;accepted: November 13, 2010.

�� 361


Documents

Functional changes in transcriptomes of the prefrontal